Pathway modeling applied to TREAT-AD proteomic targets demonstrates linkages between AD endophenotypes.

Abstract

The poor success rate of therapeutic targeting in Alzheimer's disease necessitates the explorations of new biological areas and hypotheses of disease mechanism. The TREAT-AD Consortium is an international group of academic researchers dedicated to identifying novel targets for AD from under-explored areas of disease-linked pathology. However, target identification from genomic and proteomic studies is extremely challenging. Toward that end, we identified 13 core biological domains associated with AD, and created a gene ontology (GO) mapping to each in order to automatically identify genes associated with each endophenotype. The TREAT-AD hits coming from large scale brain proteomic studies were then mapped onto the core biological domains. A novel system of dynamic data driven pathway reconstruction was applied that draws binary relationships from the Pathway Commons data store and builds AD-weighted pathways around the proteomic targets linked to the 13 core biological domains of AD. The dynamic data-driven pathway reconstruction application developed large network objects around the submitted proteomic targets that create hypothetical linkage points across discrete areas of disease biology. The network objects represent a holistic pathway linking the submitted TREAT-AD proteomic hits to each other, through missing elements and disease sentinel genes, to create an interlinked pathway model. This model can be shared and community-mined for significant AD-linked hypotheses. Using this approach, we identified numerous interactions between AD endophenotypes, such as neuroinflammation and endocytosis. One of the primary difficulties with therapeutic target identification is understanding the biological network within which any particular gene resides. The tools we have developed may help examine the biological neighborhood more robustly, as well as assess specific hypothetical mechanist connections between genes that may help identify and optimized AD-drug targets in the future.